L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1....
Transcript of L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1....
![Page 1: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/1.jpg)
L5. Quantitative population genetics
OVERVIEW . L1. Approaches to ecological modelling
. L2. Model parameterization and validation
. L3. Stochastic models of population dynamics (math)
. L4. Animal movement (math + stat)
. L5. Quantitative population genetics (math + stat)
. L6. Community ecology (stat)
![Page 2: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/2.jpg)
This nine-spined stickleback originates from the Baltic Sea population
There is variation among individuals
This nine-spined stickleback originates from Pond Pyöreälampi population
Why are they different?
![Page 3: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/3.jpg)
Ph
eno
typ
e (e
.g.,
bo
dy
size
)
Environmental effect?
Environment (e.g., amount of food)
Juha Merilä
No! We did a common garden experiment.
![Page 4: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/4.jpg)
Ph
eno
typ
e (e
.g.,
bo
dy
size
)
Environmental + genetic effect?
Environment (e.g., amount of food)
Genotype BIG
Genotype SMALL
Yes!
![Page 5: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/5.jpg)
Ph
eno
typ
e (e
.g.,
bo
dy
size
)
Environmental | genetic effect?
Environment (e.g., amount of food)
Genotype BIG
Genotype SMALL
Don’t know.
![Page 6: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/6.jpg)
Exp
ecte
d p
hen
oty
pe
(e.g
., b
od
y si
ze)
Developmental instability?
Environment (e.g., amount of food)
No! Siblings show a consistent
pattern.
![Page 7: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/7.jpg)
Why is there a genetic difference?
Charles Darwin(1809-1882)
Natural selection.
Survival of the fittest!
![Page 8: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/8.jpg)
Natural selection?
Juha Merilä
The fish are different because of local adaptation. Small fish are better in escaping predators (important in sea), big fish are better in competing for food (important in ponds).
![Page 9: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/9.jpg)
Sewall Wright (1889-1988)
Not only natural selection?
Mutation, migration and genetic drift. Adaptive landscapes.
![Page 10: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/10.jpg)
Genetic drift?
Juha Merilä
Maybe the fish are different just by chance? Could the difference be generated by the random assortment of genes from parents to offspring?
![Page 11: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/11.jpg)
Ronald Fisher (1890-1962)
Is drift a plausible hypothesis?
Most natural populations are too large for drift to be important.
![Page 12: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/12.jpg)
Is drift a plausible hypothesis?
Drift can still be important in genetically isolated sub-populations.
Theodosius Dobzhansky (1900-1975)
![Page 13: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/13.jpg)
Genetic drift?
Juha Merilä
I am studying genetically isolated sub-populations, so drift could be a plausible explanation…
![Page 14: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/14.jpg)
Russel Lande
Quantitative genetics for evolutionary biology of natural populations.
To separate drift and selection, we need quantitative tools
![Page 15: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/15.jpg)
Hire a statistician as a PhD student!
How does one apply quantitative genetics theory in practice?
Juha Merilä
Markku Karhunen
![Page 16: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/16.jpg)
Statistical methods for detecting signals of natural selection in the wild
(PhD dissertation in Helsinki, 18th October 2013)
I. Ovaskainen, Karhunen, Zheng, Cano Arias and Merilä, Genetics 2011 II. Karhunen and Ovaskainen, Genetics 2012 III. Karhunen, Merilä, Leinonen, Cano Arias and Ovaskainen, Molecular Ecology Resources 2013 IV. Karhunen, Ovaskainen, Herczeg and Merilä, Evolution (in press)
Markku Karhunen
![Page 17: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/17.jpg)
Ancestral population
Genetic differentiation by drift or selection? Ti
me
Pond population Sea population
![Page 18: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/18.jpg)
Null model: what happens under neutrality?
Related individuals resemble each other:
Cov[ , ] 2i j ija a G
A B ABCov[ , ] 2a a G
Breeding values of individuals i and j Amount of additive variance
Coancestry (relatedness) between individuals i and j
Related populations resemble each other:
MEAN breeding values in POPULATIONS A and B
MEAN coancestry (relatedness) between individuals in POPULATIONS A and B
![Page 19: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/19.jpg)
Ancestral population
Tim
e Exercise: what happens under drift?
Assume that neutral molecular data tells that in terms of coancestry the populations form 2 groups:
Past population 1 Past population 2
Current populations
![Page 20: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/20.jpg)
Trait 1 Trait 1
Trai
t 2
Exercise: what happens under drift?
CASE 1 CASE 2
Which of these patterns is more likely to have evolved due to random genetic drift?
A B ABCov[ , ] 2a a GRelated populations resemble each other:
![Page 21: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/21.jpg)
How to turn the eyeballing exercise into a statistical test?
Population-to-population relatedness matrix
Matrix of ancestral genetic variances and co-variances
Overall mean
Vector of population means
A B ABCov[ , ] 2a a GAssume and e.g. normally distributed traits. Then
S-statistic (based on Mahalanobis distance): does the observed vector of population means fit into the “core” of this distribution or is it an “outlier”?
S close to 0: stabilizing selection S close to 1: diversifying selection S close to 0.5: drift plausible
![Page 22: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/22.jpg)
How does this relate to FST-QST tests?
• FST: population divergence in neutral markers • QST: population divergence in quantitative traits
• FST<QST: diversifying selection • FST>QST: stabilizing selection • FST=QST: drift plausible
A B ABCov[ , ] 2a a G
FST
QST Unlike FST-QST tests, this equation • accounts for evolutionary stochasticity • utilizes population-specific information • extends naturally to multivariate traits
![Page 23: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/23.jpg)
Trait 1 Trait 1
Trai
t 2
Exercise: what happens under drift?
CASE 1 CASE 2
Which of these patterns is more likely to have evolved due to random genetic drift?
A B ABCov[ , ] 2a a GRelated populations resemble each other:
FST=QST S=0.5
FST=QST S>0.99
![Page 24: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/24.jpg)
Loca
l po
pu
lati
on
s an
d s
amp
led
in
div
idu
als
Lab
ora
tory
p
op
ula
tio
n
Ancestral population
Ped
igre
e B
reed
ing
exp
erim
ent
Data: neutral markers + quantitative traits
![Page 25: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/25.jpg)
Parameter estimation with a Baysian approach
Neutral marker data
Quantitative trait data from breeding experiments
Population-to-population relatedness matrix
Directed Acyclic Graph (DAG) representing a hierarchical Bayesian model
Vector of population means
Matrix of ancestral genetic variances and
co-variances
![Page 26: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/26.jpg)
Example results on nine-spine sticklebacks
BYN
BAS
PYÖ
WHS Morphological traits: S=1.00 Behavioural traits: S=0.91
S close to 0: stabilizing selection S close to 1: diversifying selection S close to 0.5: drift plausible
![Page 27: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/27.jpg)
Signature of local adaptation from habitat information?
Juha Merilä
We found small fish from BOTH sea populations, and large fish from BOTH pond populations – this must give additional evidence of selection operating.
H-statistic: are populations from similar habitats more similar than expected by
random drift?
Distance matrices of environmental covariates and mean trait values
Mantel test statistic
![Page 28: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/28.jpg)
Example results on nine-spine sticklebacks
BYN
BAS
PYÖ
WHS
Morphological traits: S=1.00 Behavioural traits: S=0.91
Morphological traits: H=1.00 Behavioural traits: H=0.99
![Page 29: L5. Quantitative population genetics - ICTP · L5. Quantitative population genetics OVERVIEW . L1. Approaches to ecological modelling . L2. Model parameterization and validation .](https://reader030.fdocuments.net/reader030/viewer/2022040410/5ecf2f86dbbbed738d71cb61/html5/thumbnails/29.jpg)
L5: take home messages
• Major forces behind evolution include selection, drift, mutation and migration
(gene-flow)
• Showing that two populations are genetically different does not necessarily
mean that the populations have been influenced by different selection
pressures.
• To rule out the alternative explanation of genetic drift, the pattern of genetic
divergence among the populations can be compared to patterns that would be
generated by drift alone.
• Quantitative genetic theory provides theoretically well-founded and statistically
powerful approaches for testing hypotheses related to selection and drift, as
well as estimating related key parameters (e.g. G-matrices, gene-flows,
coancestry matrices at the individual and population levels, etc.)